Category: Video Surveillance & Imaging

Thermal cameras detect infrared radiation emitted by objects. Because they do not rely on visible illumination, they can detect people, vehicles and equipment in darkness and many difficult lighting conditions. Their strongest role is as part of a layered detection and verification architecture.

Perimeter security

Thermal imaging is especially valuable for perimeter detection. A person may appear as a clear heat contrast when a conventional camera sees only darkness. When paired with properly configured analytics, thermal cameras can classify targets and generate alarms for operator verification.

Industrial monitoring

Thermal cameras are also used to monitor electrical equipment, batteries, process machinery and storage areas for abnormal heat. In these applications, the same sensing platform can contribute to both safety and security operations. Temperature-related conclusions require equipment designed and configured for measurement, not merely thermal video.

Performance in smoke, haze and low contrast

Thermal imaging can sometimes provide useful visibility through light smoke, haze or difficult backlighting where visible cameras struggle. Performance still depends on atmospheric conditions, target contrast, wavelength and the characteristics of the scene.

Advantages of thermal imaging

• Operation without visible illumination
• Strong contrast for many warm targets
• Useful detection capability across large or remote areas
• Potential privacy advantages where facial detail is unnecessary

Important limitations

Thermal cameras do not see through walls and cannot always identify a person. Glass may reflect or block thermal radiation depending on wavelength. Heavy rain, high humidity and extreme ambient temperatures can reduce effective detection performance. Thermal imagery also requires different interpretation from visible video.

Radiometric and non-radiometric systems

Some thermal cameras estimate temperature values, while others are intended only for imaging and detection. Temperature-measurement applications require suitable radiometric equipment, calibration, emissivity settings and environmental compensation. A non-radiometric security camera should not be treated as a precision temperature instrument.

Analytics and seasonal testing

Analytics can detect people or vehicles in thermal scenes, but training data and scene configuration matter. A system should be tested under representative seasonal and weather conditions because temperature contrast changes throughout the year.

Integration with other sensors

A common architecture uses thermal cameras for detection and visible PTZ cameras for verification. Radar can add tracking and range information. Combining complementary sensors can provide stronger situational awareness than relying on any single technology.

Choosing a thermal camera

Selection should begin with the required detection task, field of view, target distance, environmental conditions and integration workflow. Resolution, lens focal length, thermal sensitivity and environmental rating matter, but no single specification determines real-world performance.

Conclusion

Thermal imaging is a powerful specialized tool for physical security and industrial monitoring, but it is not a universal replacement for visible cameras. The best results come from risk-based design, realistic acceptance testing and integration with verification and response procedures.

A practical guide to modern video surveillance systems, covering IP cameras, PTZ, thermal imaging, VMS, storage, analytics, interoperability and system design.

What is a video surveillance system?

A video surveillance system captures visual information from a protected environment and makes that information available for live monitoring, recording, search, investigation and increasingly automated analysis. Traditional analog CCTV systems carried video over dedicated coaxial infrastructure. IP-based systems encode video digitally and transmit it across networks, which makes integration, remote access and distributed architectures much easier.

The main components are cameras, network infrastructure, recording or cloud storage, a video management system, operator interfaces and optional analytics. In larger installations, identity systems, intercoms, intrusion sensors and command-and-control platforms may also exchange events with the video system.

Camera types and where they fit

Fixed network cameras are suited to entrances, corridors and areas where the required field of view is known. PTZ cameras can pan, tilt and zoom, making them useful for active tracking and wide-area verification. Multi-sensor and panoramic cameras reduce blind spots by covering several directions from one mounting point. Thermal cameras detect infrared radiation rather than visible light, so they can be useful for detection in darkness, haze or visually complex perimeter environments. They are not a universal replacement for visible cameras because identification and forensic detail may still require conventional imaging.

Resolution is only one part of image quality. Lens selection, field of view, lighting, shutter settings, dynamic range, compression and mounting position can matter as much as pixel count.

VMS: the operating layer

Video Management Software, or VMS, is the layer that brings camera streams, recordings, alarms, maps, user permissions and investigations together. It controls who can see what, how video is recorded and retained, how operators receive alarms and how recorded material is searched or exported.

Interoperability is important in multi-vendor systems. ONVIF describes itself as an open industry forum for standardized interfaces for IP-based physical security, covering advanced video streaming, edge storage and metadata/events for analytics applications. Conformance should be checked against the official ONVIF conformant-products database rather than assumed from a marketing claim.

Storage and bandwidth

Storage requirements depend on resolution, frame rate, compression, scene complexity, recording mode and retention period. H.265 can reduce bandwidth and storage compared with older encoding approaches under suitable conditions, but real-world savings vary. Motion-based recording can reduce storage in low-activity scenes, while high-activity environments may record almost continuously.

For critical systems, storage design should also consider redundancy, evidence integrity, cybersecurity and what happens when network connectivity is lost.

Where AI analytics fit

Video analytics can detect or classify events such as people, vehicles, line crossing, loitering or objects entering a restricted area. The useful question is not whether a camera has “AI,” but how accurately the analytic performs in the actual scene and how false alarms are handled. Analytics should be treated as decision support, since lighting, weather, camera angle and occlusion can all affect performance.

How to specify a system

Start with threats and operational goals. Define what must be detected, what level of identification is needed, how quickly an alarm must be verified, which areas are most critical and how long video must be retained. Then design fields of view, pixel density, lighting, network capacity, storage, resilience and operator workflow around those requirements.

FAQ

Is IP video always better than analog CCTV? IP architectures offer major advantages in integration, scalability and analytics, but migration decisions should consider existing infrastructure, cybersecurity, cost and operational need.

What does ONVIF compliance mean? It means a product conforms to a defined ONVIF profile. Compatibility should be verified for the specific profiles and functions required by the project.

Do AI cameras eliminate human monitoring? No. They can reduce search and alarm workload, but human verification remains important, especially in high-consequence environments.

Verification note: Avoid presenting a vendor’s claimed AI accuracy as a universal performance figure. Test results are scene- and configuration-dependent.